Figure 1. Humanizing
tdp-1 to determine if patient variants cause stress-induced neurodegeneration:A. Strategy for genome editing. A wild type (N2, white) animal is CRISPR-edited to make a knock-out (KO, red) animal. When loss-of-function defects are detected in the KO, the gene is humanized (blue); wild type C. elegans exons and introns are excised and replaced with human coding sequences. Finally, another round of CRISPR-editing is undertaken to insert clinical variants (grey) into the humanized locus. Comparison of the N2 and KO permits detection of loss of function defects. Comparison of these with the humanized-control reveals functional orthology by rescue, and comparison to patient-associated allele lines allows detection of variant defects.B. Knock out of
tdp-1 locus. Creation of a
tdp-1 KO null allele by CRISPR-editing required deleting sequences (red) containing exons and introns (grey boxes and intervening lines).C. Human gene replacement at
tdp-1 locus. CRISPR-editing was used to remove endogenous coding sequences and insert human gene coding sequences to create hTADRBP, containing artificial introns (blue exon boxes linked by orange artificial intron line).D. Cis editing predominates. A CRISPR-editing approach was used to insert hTARDBP sequences into the
tdp-1 locus (black), which is closely linked to the
dpy-10 locus (grey). These loci are ~2 centiMorgans apart. At the co-CRISPR marker
dpy-10 locus, one allele was converted to
dpy-10(
cn64) and the other allele was unedited (hemi-converted); the resulting Roller heterozygotes were selected (green). Co-conversion of the
tdp-1 target locus with hTARDBP fragment insertion occurred only on the cis strand (blue).E. Iterative assembly using linked alleles. Target genes that are tightly linked to a coCRISPR marker can be iteratively co-CRISPR-edited to build large transgenes by flip-flop selection procedure. In this example, the
dpy-10 locus is co-CRISPR converted to yield
dpy-10(
cn64) (green) and conversion of the linked locus to contain a gene fragment content (dark blue) in the first step. In the second step, using the Dumpy animals from the first step, a repeat round of co-CRISPR restores
dpy-10(wt) (grey) simultaneously with insertion of another gene fragment (light blue). Repeat rounds of co-CRISPR using the Dumpy and nonDumpy phenotypes allows iterative editing of the target locus and efficient sequential build up of a large gene assembly (additional blue boxes).F. Assessing neurodegeneration C. elegans phasmid neurons take up fluorescent lipophilic dyes via exposed sensory endings. Sensory ending defects, process retraction or neuron loss can cause dye-filling defects.G. Mild stress with 2.5mM paraquat for 22 hours does not cause dye filling defects in N2 (white), WT hTARDBP (HA3971, blue), or
tdp-1(-) animals (lof, HA3703). Three independent trials; 20 animals per genotype per trial.H. Moderate stress with 10mM paraquat for 22 hours causes dye filling defects in
tdp-1(
tgx58) loss of function animals (lof, HA3703, red) that exceed defects observed in N2 animals (white) or WT hTARDBP animals (HA3971, blue). Twelve independent trials; 20 animals per genotype per trial. By student's T-test, p value for
tdp-1(lf) vs N2 is 0.001 and vs hTARDBP is 0.056, while hTARDBP vs N2 is likely not significant at 0.11.I. Patient variants inserted into hTARDBP Generation of patient variants M337V, A315T, G298S, and G294A did not result in moderate stress-induced dye filling defects, only G295S differed from hTARDBP. Three independent trials with 20 animals per genotype per trial; hash marks on X-axis group separate genotypes assayed simultaneously. Note that the experimental results shown in Panel G for N2, hTARDBP and
tdp-1(-) are the same experimental results as shown in Panel F; results are consolidated in Panel F for ease of presentation. Strains used were HA4008 for M337V, HA4006 for A315T#1, HA4007 for A315T#2, HA4003 for G298S, HA4005 for G294A, HA3983 for G295S. By student's T-test, p value for WT versus G298S is <0.05.